4 Lunn Article 48-1

The Economic and Social Review, Vol. 48, No. 1, Spring, 2017, pp. 85-107 Are Perceptions of Greatness Accurate? A Statistical Analysis of Brian O’Driscoll’s Contribution to the Irish Rugby Team Peter D. Lunn* The Economic and Social Research Institute and Trinity College Dublin David Duffy† Property Industry Ireland Abstract: We test whether public perceptions of Brian O’Driscoll’s greatness were accurate. O’Driscoll was an automatic selection for Ireland’s rugby team yet missed matches through injury, allowing us to treat injury as random. We estimate that O’Driscoll was worth a converted try per game and improved the chance of victory by more than home advantage. None of O’Driscoll’s Irish contemporaries made such a contribution. His impact also compares favourably with New Zealand’s best players of the era. Our estimates have implications for debates about wages in professional rugby. They also demonstrate how statistical techniques can be applied to an everyday issue, allowing us to conclude that perceptions of one Irishman’s greatness were warranted. I INTRODUCTION y common consent, Brian O’Driscoll is one of Ireland’s sporting greats. BMeasured in terms of caps, tries or trophies, his playing record is unsurpassed in the era of professional rugby. On retirement, O’Driscoll had broken the international record number of caps for an Irish player, scored the highest number tries for Ireland, and scored more tries than any other centre in world rugby. The public perception of O’Driscoll is that he was exceptional among all of his peers. In his latter years, perhaps having run out of adjectives, multiple members of the Irish press began referring to him simply as “the great one”. Acknowledgements: We thank David Byrne, Alan Barrett, Kieran McQuinn and an anonymous referee for helpful comments and feedback on earlier drafts. Since one author is a Dublin-based Englishman who has adopted Leinster, and the other is Irish and a Munster fan, any conflicts of interest have been internalised. * Corresponding author: [email protected] † This work was completed while David was a Senior Research Officer at the ESRI. 85 86 The Economic and Social Review Are these public perceptions biased or accurate? Just how good was Brian O’Driscoll? Among many Ireland fans, he was simply the greatest Irish rugby player of all time and on a par with any other player on the world stage. Yet, while his performances earned him international admiration and respect, many non-Irish rugby fans thought that Ireland fans were inclined to exaggerate about O’Driscoll’s brilliance, especially later in his career.1 This is, of course, entirely plausible. In addition to the potential for bias due to national allegiance, several other biases documented in the experimental psychology literature might generate a tendency to inflate greatness. There is evidence that first impressions can colour later judgements, producing a so-called “halo effect” (Thorndike, 1920). Applied to sport, this effect implies that where a player bursts onto the scene with eye-catching performances, their subsequent contributions are likely to be judged too positively. Brian O’Driscoll captured widespread attention in his first international competitive tournament, the Six Nations of 2000, when he scored a hat-trick of tries in a victory against France in Paris. It is possible that such early success coloured later perceptions. A related phenomenon is “confirmation bias”, whereby ambiguous new evidence tends to be interpreted in line with prior beliefs (Nickerson, 1998). Thus, errors by a sportsperson believed to be of exceptional standard may be put down to extraneous circumstances rather than poor performance, while fortunate turns of events may be interpreted as further signs of sporting genius. The present paper seeks an objective view. We take advantage of unusual properties of Brian O’Driscoll’s career, which allow us to avoid several of the usual pitfalls that afflict the measurement of sporting performance. We estimate O’Driscoll’s individual impact on three outcomes in relation to each match: the points difference, whether the match was won and the scale of victory or defeat (more or less than one score, i.e., seven points). For comparison, we also perform an equivalent analysis for two New Zealand players who are considered to be exceptional on the international stage during the same era: Dan Carter and Richie McCaw. The primary purpose is to test whether public perceptions of greatness, at least in this case, are prone to exaggeration. A second aim is to contribute to the ongoing debate, in Ireland especially, about the economic value of rugby players in the professional era. Modern rugby faces difficult economic issues regarding how best to allocate revenues generated from gate receipts and television rights. It must find a balance between the freedom of players to earn due rewards for their skills, the desire for successful national teams, and the broader development of the game, in a context where larger and smaller nations tend to differ on the optimum allocation. Rugby authorities in some nations have capped wages and restricted the labour 1 In the Northern Hemisphere see, for instance, “The Rolling Maul: pray for a rejuvenated Brian O’Driscoll”, Stephen Jones, The Times, 27 May, 2009. In the Southern Hemisphere, there is a lively debate among New Zealand rugby fans as to how many All Blacks caps O’Driscoll would have got, with some arguing that he would have struggled to make their team (www.reddit.com/r/rugbyunion). A Statistical Analysis of Brian O’Driscoll’s Contribution to the Irish Rugby Team 87 market mobility of players wishing to represent their countries. In this context, we provide what can be regarded as an estimate of the upper limit on the contribution of a single player to the team outcome. Lastly, the paper demonstrates how econometric techniques can be applied to everyday matters. Some may wish to argue that greatness, in sport and other walks of life, is and always will be a largely subjective concept, beyond the reach of statistics. Here, we employ data gathered from publicly available sources to show that statistics can sometimes produce objective evidence even about such apparently subjective matters. In sport, players are generally judged on their record. O’Driscoll won his first Ireland cap at age 20 in June 1999, on a tour of Australia. He was capped 133 times for Ireland, including 83 as captain (2003-2012) and four World Cups. He also won eight caps for the British and Irish Lions. He scored 46 tries for Ireland; the eighth- highest number in international rugby union history and the highest ever for a centre. Ireland’s primary competition is the Six Nations Championship. While O’Driscoll was captain, Ireland won the Triple Crown (beating England, Wales and Scotland in the same year) in 2004 (Ireland’s first since 1985), 2006 and 2007. In 2009, Ireland won the Triple Crown, the Six Nations Championship and their first Grand Slam (beating all five opponents) in 61 years. O’Driscoll was Player of the Tournament in 2006, 2007 and 2009. Considered overall, while this is a remarkable record and these are impressive facts, many of these statistics depend, in part, on factors other than individual performance: players can only win caps and score tries if selected; the likelihood of winning depends on the opposition, location and timing of matches; awards are made to individuals by subjective judgment; teams not individuals win trophies. Thus, O’Driscoll’s record, while impressive, does not necessarily provide a reliable measure of his individual contribution. There is an expanding academic literature, outlined briefly in the next section, that seeks more accurate measures of individual performance across sports. But accurate assessments of the contributions of individual players to team results generally elude academic researchers, because it is extremely difficult to quantify the contribution of a single player to the team outcome. An ideal test would compare the result of games, on average, when the individual player was in the team, with the average result when the specific individual was not in the team, all else equal. This would indicate the impact of the player on team outcomes, at least compared to whatever alternative player was available to fill their boots. Unfortunately, however, simply comparing results when the player plays and when the player does not can give an inaccurate estimate; all else is rarely equal. Even if statistical techniques are used to control for other factors likely to influence the result (whether the game is home or away, who the opposition are, when it is played, and so on), factors that influence selection for the team in the first place also influence the estimate. Selection is generally the decision of the coach (or manager). Players are sometimes rested or rotated. Players are less likely to be selected early in their career, when trying to establish themselves, and later in their 88 The Economic and Social Review career, as their athletic abilities decline. Selection partly depends on current form and fitness, both of the player and others competing for the same shirt. Thus, differences between average outcomes when the player is and is not in the team depend, not simply on the contribution of the individual player, but also on selection decisions. In short, whether the player takes the field or not is endogenous. In this context, however, Brian O’Driscoll’s career is highly unusual, because at all times during his 15-year international career O’Driscoll was an automatic choice. Although he was controversially dropped from the British and Irish Lions team for the last test of his final tour in 2013, O’Driscoll’s selection for Ireland was never questioned for internationals against top opposition.2 When Ireland played in the Six Nations Championship, or against frontline opposition from the Southern Hemisphere, O’Driscoll was in the Ireland XV provided he was fit to play.

Load more